Nuclear Factor of Activated T Cells (NFAT) Proteins Repress Canonical Wnt Signaling via Its Interaction with Dishevelled (Dvl) Protein and Participate in Regulating Neural Progenitor Cell Proliferation and Differentiation*

The Ca2+ signaling pathway appears to regulate the processes of the early development through its antagonism of canonical Wnt/β-catenin signaling pathway. However, the underlying mechanism is still poorly understood. Here, we show that nuclear factor of activated T cells (NFAT), a component of Ca2+ signaling, interacts directly with Dishevelled (Dvl) in a Ca2+-dependent manner. A dominant negative form of NFAT rescued the inhibition of the Wnt/β-catenin pathway triggered by the Ca2+ signal. NFAT functioned downstream of β-catenin without interfering with its stability, but influencing the interaction of β-catenin with Dvl by its competitively binding to Dvl. Furthermore, we demonstrate that NFAT is a regulator in the proliferation and differentiation of neural progenitor cells by modulating canonical Wnt/β-catenin signaling pathway in the neural tube of chick embryo. Our findings suggest that NFAT negatively regulates canonical Wnt/β-catenin signaling by binding to Dvl, thereby participating in vertebrate neurogenesis.

Dishevelled (Dvl) is an essential effector in this pathway. Despite its function on ␤-catenin accumulation, it also plays a role at ␤-catenin transcriptional activity. Previous studies have demonstrated that Dvl is also localized in the nucleus (9 -11), and its nuclear localization is required for canonical Wnt signaling (10,11). Nuclear Dvl associates with the ␤-catenin-LEF1/TCF transcriptional complex by interacting with ␤-catenin and c-Jun to stabilize ␤-catenin-LEF1/TCF transcriptional complex formation (11).
Proteins of the nuclear factor of activated T cells (NFAT) family are Ca 2ϩ -regulated transcription factors, which were initially identified in the immune system but now are known to have integral function in multiple signaling pathways (12)(13)(14)(15)(16). Five members of this family, including NFATc1/2/c, NFATc2/ 1/p, NFATc3/4/x, NFATc4/3, and NFAT5/TonEBP, have been identified in mammals. All members of the NFAT family have highly conserved DNA-binding domain, the Rel homology region (RHR), which confers common DNA-binding specificity (17). Except for NFAT5, NFATc1-c4 are activated upon a rise in intracellular Ca 2ϩ , which stimulates the serine/threonine phosphatase activity of calcineurin (18,19). Calcineurin directly dephosphorylates NFAT resulting in the nuclear import of NFAT. In the nucleus, NFAT transcription complexes reassemble on DNA to initiate gene transcription (19). It has been revealed that NFAT can act as both a positive and negative regulator of transcription. A study of cyclin-dependent kinase 4 (CDK4) suggested that NFATc2 represses CDK4 expression (20). NFATc1 also induces transcriptional repression by interacting with histone deacetylases after sumoylation modification (21,22).
Previously, it has been reported that Ca 2ϩ signaling or Wnt/Ca 2ϩ signaling can block the canonical Wnt signaling transduction through an activation of NFAT in Xenopus embryos (23). However, the underlying mechanism is not clear. Here, we report that NFAT participates in the inhibition of Wnt signaling through direct interaction with Dvl in the nucleus. The binding of NFAT to Dvl prevents Dvl recruiting to the ␤-catenin transcription complex, thus reducing the transcriptional activity of ␤-catenin. Furthermore, our data also reveal that the cross-talk between NFAT and Wnt signaling is involved in the proliferation and differentiation of neural progenitor cells in the neural tube of chick embryo.
Cell Transfection and Luciferase Assay-HEK293T were transfected with DNA using Lipofectamine Plus (Invitrogen). SW480, Caco-2, and LS174T cells were transfected with DNA using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. The LacZ plasmid was added to make the total amount of DNA equal (0.5 g/well in a 24-well plate for SW480 cells and 0.25 g/well for HEK293T cells). Luciferase assays were performed as described previously (25).
Immunoprecipitation, Protein Purification, and in Vitro Pulldown Assay-HEK293T cells were transiently transfected with the indicated constructs for 24 h. A co-immunoprecipitation experiment was then performed as described previously (25). His 6 -tagged and GST-fused proteins were expressed in Escherichia coli and were partially purified with nickel-nitrilotriacetic acid and GSH-Sepharose 4B beads, respectively. The in vitro pulldown assay was performed as described previously (11).
Cell Fractionations and Endogenous Interaction-HEK293T cells were scraped into PBS and were fractionationed as described previously (26). For the endogenous interaction assay, the nucleus of 3-5 ϫ 10 7 HEK293T cells were lysed, and the nuclear extracts were sonicated four times for 5 s in 1 ml of buffer containing 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100, 2 mM EDTA, and Complete TM protease inhibitors (Roche Applied Science).
In Ovo Chick Embryo Electroporation-Fertilized eggs were obtained from the Shanghai Academy of Agricultural Sciences. The chick embryo in ovo electroporation was performed as described previously (28).
In Situ Hybridization-Section in situ hybridization was performed as described previously (29).

RESULTS AND DISCUSSION
Identification of NFAT as a Dvl-binding Protein-The Dvl protein family is an essential yet also perhaps the most enigmatic component of the Wnt signaling pathway. To address its function further, we sought to identify proteins acting directly with Dvl. A yeast two-hybrid screen of a mouse embryonic cDNA library was performed using mouse Dvl-1-C1 (residues 377-695) as the bait. A positive clone that encodes the C-terminal half of mouse NFATc4 (mNFATc4-C1) interacted strongly with the bait. To confirm this interaction, we carried out co-immunoprecipitation after transfecting HEK293T cells with HA-tagged mNFATc4 and FLAG-tagged Dvl-1, 2, and 3. All three Dvl homologs were co-immunoprecipitated with HAtagged mNFATc4 (Fig. 1A), which suggests that NFAT binds to Dvl and has a universal binding activity to Dvl family members. We further purified His 6 -tagged full-length mNFATc4 and GST-mDvl-1 from E. coli. GST-mDvl-1, not GST alone, pulled down recombinant His 6 -tagged mNFATc4 in vitro (Fig. 1B), suggesting that NFAT can physically interact with Dvl. Moreover, we delineated the binding region within mNFATc4. As described in Fig. 1C, the N terminus of NFAT composed of the transactivation domain and regulatory domain is dispensable for interacting with Dvl, whereas the RHR has binding activity slightly less than that of full-length NFAT. Thus, we consider the RHR as the region critical for binding to Dvl.
Proteins of the NFAT family play important roles in the Ca 2ϩ signaling which induces dephosphorylation of NFAT leading to its nuclear transport. Thus, we examined whether calcium affected the interaction of NFAT and Dvl. We used the calcium ionophore ionomycin to mimic the Ca 2ϩ signaling (30 -33). Because the available NFAT-specific antibody does not detect endogenous NFAT well, we stably expressed FLAG-tagged mNFATc4 in HEK293T cells. NFAT was highly phosphorylated upon the stimulation with CsA, a calcineurin inhibitor, and was dephosphorylated after ionomycin was added (supplemental Fig. S1A), indicating that the cells responded to calcium signaling. We then carried out immunoprecipitation to detect the interaction of endogenous Dvl-3 with NFAT upon CsA or ionomycin treatment. As shown in Fig. 1D, Dvl-3 efficiently co-immunoprecipitated with NFAT in the presence of ionomycin, but did so weakly upon CsA treatment. To confirm further whether the two different forms of NFAT caused by CsA and ionomycin treatment have different affinity for Dvl, we stimulated the cells stably expressing NFAT with CsA or ionomycin and carried out immunoprecipitation with a FLAG antibody. Then, Dvl expressed in HEK293T cells was added. As shown in Fig. 1E, NFAT from Ca 2ϩ -stimulated samples appeared to have a higher binding affinity for Dvl. Given that the dephosphorylated form of NFAT is located predominantly in the nucleus. We further performed a co-IP experiment using an extract of the nucleus fraction to verify whether the interaction happens in the nucleus. Indeed, NFAT forms complex with Dvl in the nucleus upon the stimulation of Ca 2ϩ signal (Fig. 1F). These data suggest that the interaction between NFAT and Dvl possibly takes place in the nucleus upon Ca 2ϩ concentration elevation.
NFAT Negatively Regulates Wnt Signaling-It has been reported that Ca 2ϩ or Wnt/Ca 2ϩ signaling plays a negative role in regulating Wnt/␤-catenin signaling (23). Thus, we used the LEF1/TCF-dependent reporter TOP-Flash to validate its effect in our system. As shown in Fig. 2A, ionomycin or expression of ⌬CnA inhibited reporter gene activity induced by Wnt3a CM. This inhibition was reversed by a dominant negative form of mNFATc4 (DN-NFAT, mNFATc4-N in Fig. 1C) which only contains the N-terminal region of mNFATc4 and blocks endogenous NFAT activation (34 -36) (supplemental Fig. S1B). However, DN-NFAT could not reverse the inhibitory effect caused by a constitutively active form of mNFATc4 (mNFATc4-CA) (Fig. 2C). mNFATc4-CA lacks its phosphorylation sites (sp1, sRR1/2) and is constitutively localized in the nucleus (supplemental Fig.  S1C). It also interacts with Dvl (supplemental Fig. S1D). We also examined the effects of calcineurin or DN-NFAT expression on the mRNA levels of Axin-2 and DKK1, specific Wnt target genes ( Fig. 2D and supplemental Fig. S2A), which provides a coincident result as reporter gene assay. These data confirmed the previous findings that NFAT is not only a component of calcium signaling (Fig. 2D, right), but also a regulator that mediates the inhibition of the Wnt pathway by calcium signaling, especially via calcineurin.
NFAT proteins can also act as transcriptional activator. We investigated whether the transcriptional activity of NFAT was required for its inhibitory effect on Wnt signaling. The RHR domain of mNFATc4 contains the DNA-binding activity but not transactivation activity (supplemental Fig. S2B). The fact that RHR alone could suppress Wnt signaling (Fig. 2E) suggests that the transcriptional activity of NFAT is dispensable for its ability to suppress Wnt signaling.
NFAT Functions Downstream of ␤-Catenin-It has been reported previously that NFAT participates in the regulation of Wnt signaling by acting upstream of ␤-catenin (23). However, in our system, neither calcineurin expression nor NFAT-CA expression interfered with the contents of phosphorylated ␤-catenin (Ser-33, Ser-37, and Thr-41) or ␤-catenin nuclear accumulation whereas expression of GSK3␤ or Axin did (Fig. 3,  A and B). These results indicate that NFAT may not influence the stability of ␤-catenin. Therefore, we postulated that NFAT may repress Wnt signaling in the nucleus by blocking the transactivation function of ␤-catenin. This hypothesis is consistent with the aforementioned finding that the interaction of NFAT with Dvl is induced by Ca 2ϩ signals in the nucleus. To test this hypothesis, LS174T cells that contain oncogenic mutation of ␤-catenin gene (37,38) and Caco-2 or SW480 cells which contain mutated adenomatous polyposis coli (APC) (39) were used to examine the effect of Ca 2ϩ signaling/NFAT on Wnt reporter activity. As shown in Fig. 3C, overexpression of calcineurin or NFAT-CA led to a repression of Wnt reporter gene activity in these cells. Taken together, these data suggest that calcineurin and NFAT regulate Wnt signaling downstream of ␤-catenin. The numbers indicate amino acids. The binding of full-length NFAT and various NFAT fragment to Dvl was examined by co-immunoprecipitation. TAD, transactivation domain. Ϫ, no interaction; ϩ, strength of interaction. D, endogenous interaction occurred between NFAT and Dvl-3 upon stimulation with CsA or ionomycin. mNFATc4-FLAG stable cells were stimulated with 1 g/ml CsA or 1 g/ml ionomycin for 8 h. A co-immunoprecipitation assay was carried out with an anti-Dvl-3 antibody. E, difference in the interaction with Dvl between two different forms of NFAT is shown. NFAT stable cells were stimulated as indicated, and then the immunoprecipitation was carried out with anti-FLAG antibody. 2 h later, the Dvl expressed in 293T cells was added for another 1 h. F, endogenous interaction between NFAT and Dvl-3 in the nucleus. mNFATc4-Flag stable cells were stimulated with 1 g/ml CsA or 1 g/ml ionomycin for 8 h. Nuclear extracts were lysed for the co-IP experiment with anti-FLAG antibody.
Saneyoshi et al. reported that NFAT functions upstream of ␤-catenin when co-expressed with ␤-catenin in Xenopus (23). We believed that this difference may be caused by using different system or different detection approach.
NFAT Physically Influences the Interaction between Dvl and ␤-Catenin-We further delineated the regions within Dvl responsible for its interaction with NFAT to residues 450 -695, which is the same region that binds to ␤-catenin (Fig. 4A). The interaction of Dvl with ␤-catenin was shown to enhance the stability of ␤-catenin-TCF complex (11). Thus, we suspected that NFAT might disrupt the interaction between Dvl and ␤-catenin to inhibit Wnt/␤-catenin signaling. We purified His 6 -tagged ␤-catenin, His 6 -tagged NFAT-FLAG, and GST-mDvl-1 from E. coli. In the competition binding assay, NFAT disrupted GST-mDvl-1 and His 6 -␤-catenin interaction in a dose-dependent manner (Fig. 4B). Furthermore, the endogenous interaction of Dvl and ␤-catenin in the nucleus under Wnt stimulation could also be disrupted by overexpression of NFAT-CA and ⌬CnA (Fig. 4C). These data together suggest that NFAT inhibits Wnt signaling by interfering with the formation of stable Dvl-␤-catenin complex. Given that the interaction of Dvl and ␤-catenin is important for the transcription complex formation on the Wnt target gene promoter (11), we used ChIP to determine the effect of NFAT on the binding of ␤-catenin and Dvl to the Axin-2 (Fig. 4D) and the c-myc promoters (supplemental Fig. S2C). Overexpression of ⌬CnA or  NFAT-CA decreased the amounts of Dvl and ␤-catenin on the Axin-2 promoters, whereas TCF4 bound to the promoters remained unaffected (Fig. 4D). Moreover, DN-NFAT could rescue the inhibition effect of ⌬CnA (Fig. 4D). These results support a conclusion that NFAT functions downstream of ␤-catenin probably by disrupting the interaction of ␤-catenin with Dvl.

NFAT Antagonizes Wnt Signal in Regulating the Proliferation and Differentiation of the Neural Progenitor Cells in the Neural
Tube of Chick Embryo-To investigate the physiological role of NFAT in modulating canonical Wnt signaling, we extended our analysis to the central nerve system development in chick embryos. Wnt proteins are important morphogens during the development of the CNS. Wnt signaling through activation of the ␤-catenin/TCF pathway promotes the proliferation of neural precursor cells and prevents differentiation (40). We sought to determine whether NFAT plays a role in this process via its regulation of the Wnt/␤-catenin signaling. First, we performed in situ hybridization analysis to detect the expression of three predicted chick NFAT genes, NFATc1, c2, and c3 in the developing spinal cord. As shown in supplemental Fig. S3, NFATc1 is expressed predominantly in the neural tube among three NFATs in chick spinal cord. Interestingly, the expression of NFATc1 is restricted in the ventricular zone during development, and it is highly expressed in the late stages, where Wnt activity must be limited for progenitor cell differentiation (40). Second, we electroporated a NFAT reporter pNFAT-RE-TK-RFP into the developing neural tube of chicken embryos at Hamburger and Hamilton (HH) stages 10 and 11 (41) to detect the active form of endogenous NFAT. The expression of RFP was controlled by four repeats of the NFAT responsible element (NFAT-RE) from NFAT-luciferase reporter and was analyzed at HH stages 17 and 18 to visualize the endogenous NFAT activity. As shown in Fig. 5A, RFP was expressed across the neural tube from the dorsal to the ventral side, which indicates that endogenous NFAT is activated during the course of development. Moreover, we tested whether NFATc1 could also interact with Dvl. We cloned both human and mouse NFATc1, respectively, and the immunoprecipitation was carried out. As shown in Fig. 5B, hNFATc1 and mNFATc1 were co-immunoprecipitated with Dvl as well as mNFATc4. Given that the RHR domain in NFAT is responsible for binding to Dvl and it is conserved among species, we assume that chick NFAT protein could also interact with Dvl.
We further co-electroporated the pTOP-Flash reporter with DN-NFAT to test whether endogenous NFAT would modulate Wnt signaling in the neural tube. Overexpression of DN-NFAT which inhibits endogenous NFAT activity (Fig. 5A) increased the luciferase activity of pTOP-Flash (Fig. 5C), which suggests an inhibitory effect of NFAT in Wnt signaling. Because chick NFATc1 is expressed predominantly in the neural tube (supplemental Fig. S3), we performed RNAi experiment to knock down endogenous cNFATc1 (supplemental Fig. S4A), which led to increase in the activity of TOP-Flash (supplemental Fig. S4B). We also used pTOP-TK-RFP to mark the range of activity of the Wnt signaling. RFP expression was also up-regulated by DN-NFAT and spread throughout the whole neural tube, with a disturbed Wnt gradient along the dorsal-ventral axes (Fig. 5, D  and E). Taken together, these results indicate that endogenous NFAT activity also negatively regulates Wnt signaling in chick neural tube.
During CNS development, the canonical Wnt pathway functions as a mitogenic signal of neural progenitor cells (40). Activation of the Wnt/␤-catenin pathway blocks neural precursors differentiation (42). Thus, we hypothesized that NFAT may participate in regulating this process by interfering with Wnt signaling. Immunostaining of Tuj1, a neuron-specific class III ␤-tubulin, was used to measure the number of mature neurons at HH stages 20 and 21. Overexpression of DN-NFAT significantly decreased Tuj1-positive cells on the transfected side compared with the control side, whereas transfection of a GFP plasmid had no effect (Fig. 5, F and G, left). In contrast, BrdUrd staining, which monitors the proliferation of neural progenitor cells, was increased by DN-NFAT (Fig. 5, F and G, right). These data indicate that endogenous Ca 2ϩ /NFAT signaling negatively regulates neural progenitor cell proliferation and positively regulates differentiation. To test whether this function depends on its inhibition of Wnt signaling, we co-electroporated with ⌬N-LEF1 to repress Wnt signaling through binding to the TBE site (Fig. 5, F and G). Consistent with our hypothesis, co-expression of ⌬N-LEF1 rescued the decrease of Tuj1-positive cell number as well as the increase of BrdUrd cell numbers (Fig. 5, F  and G). Taken together, the above data suggest that NFAT pathway regulates neural progenitor cell proliferation and differentiation in the neural tube of chick embryo by antagonizing Wnt signaling. We noticed that Graef and colleagues employed FIGURE 4. NFAT disrupts the interaction between Dvl and ␤-catenin. A, schematic represents mDvl-1 fragments. The interaction of Dvl-1 with ␤-catenin or NFAT was detected by co-immunoprecipitation. B, NFAT competes with ␤-catenin in binding to Dvl-1 in an in vitro binding assay. His 6 -␤catenin was mixed with GST-mDvl-1 or GST-bound GSH-Sepharose 4B with or without His 6 -mNFATc4-FLAG for 3 h. Samples were subjected to Western blotting. C, NFAT-CA and ⌬CnA reduce the endogenous interaction of Dvl-3 and ␤-catenin in the nucleus. HEK293T cells were treated with Wnt3a CM for 3 h. Nuclear extracts were lysed for the co-immunoprecipitation experiment with anti-Dvl-3 antibody. D, ⌬CnA and NFAT-CA affect the accumulation of ␤-catenin and Dvl-3 on the promoter of the Axin2 gene. HEK293T cells were treated with Wnt3a CM for 2 h, and ChIP-quantitative PCR analyses were performed with anti-␤-catenin, anti-Dvl-3, and anti-TCF4 antibodies and normalized by input. The activity of LacZ was set as 100%. The data are expressed as the mean Ϯ S.E. (error bars) of three independent experiments. *, p Ͻ 0.05; ***, p Ͻ 0.001. OCTOBER 28, 2011 • VOLUME 286 • NUMBER 43 triple mutant mice to illustrate that the NFATc2/c3/c4 mutant did not interfere with ␤III-tubulin or certain cell-specific markers in the CNS (43). In our system, chick NFATc1, among the three chick NFAT types, was expressed predominantly in the neural tube. Furthermore, we used a dominant negative form of NFAT to block the activation of all three NFAT types. Thus, as a result, we observed the function of NFAT in the regulation of neural differentiation. The negative regulation of Wnt signaling by NFAT allows the progenitor cell to exit the cell cycle and differentiate (Fig. 5, F and G). However, the upstream stimulation of NFAT is not clear yet. The Ca 2ϩ -calcineurin signal is the most important signal for regulating NFAT activation, but the signal that leads to Ca 2ϩ influx during neural tube differentiation is still unclear.

NFAT Modulates Wnt Signaling via Its Interaction with Dvl
In summary, we propose a framework in which NFAT modulates canonical Wnt signaling based on this work and our previous finding (11) (Fig. 5H). NFAT negatively regulates the Wnt signaling through interacting with Dvl, which disrupts the association of Dvl and ␤-catenin, leading to the destabilization of ␤-catenin-LEF1/TCF transcriptional complex. Furthermore, NFAT functions as a "modulator" of neural progenitor cell differentiation. In the neural tube of chick embryos, endogenous NFAT is progressively expressed in the ventricular zone of neural tube at the time when the Wnt signaling needs to be reduced. The suppression of the Wnt signaling induces precursor cells to exit the cell cycle and promotes their differentiation, which comprises the function of NFAT in the development of the CNS.